Department of Clinical Epidemiology, Aarhus University Hospital, Aarhus N, Denmark.
Department of Orthopedic Surgery, Regional Hospital Horsens, Horsens, Denmark.
Acta Orthop. 2021 Apr;92(2):215-221. doi: 10.1080/17453674.2020.1863688. Epub 2020 Dec 18.
Background and purpose - Understanding the key drivers of hospital variation in postoperative infections after hip fracture surgery is important for directing quality improvements. Therefore, we investigated variation in the risk of any infection, and subgroups of infections including pneumonia and sepsis after hip fracture surgery.Methods - In this nationwide population-based cohort study, all Danish patients aged ≥ 65 undergoing surgery for an incident hip fracture from 2012 to 2017 were included. Risk of postoperative infections, based on data from hospital registration (hospital-treated infections) and antibiotic dispensing (community-treated infections), were calculated using multilevel Poisson regression analysis. Hospital variation was evaluated by intra-class coefficient (ICC) and median risk ratio (MRR).Results - The risk of hospital-treated infection was 15%. The risk of community-treated infection was 24%. The adjusted risk varied between hospitals from 7.8-25% for hospital-treated infection and 16-34% for community-treated infection. The ICC indicated that 19% of the adjusted variance was due to hospital level for hospital-treated infection. The ICC for community-treated infections was 13%. The MRR showed a 2-fold increased risk for the average patient acquiring a hospital-treated infection at the highest risk hospital compared with the lowest risk hospital. For community-treated infection, the MRR was 1.4.Interpretation - Our results suggest that 20% of infections could be reduced by applying the top performing hospitals' approach. Nearly a 5th of the variation was at the hospital level. This suggests a more standardized approach to avoid postoperative infection after hip fracture surgery.Hip fracture is a leading cause of hospital admission among the elderly. The 30-day mortality following hip fracture surgery has been approximately 10% during the last few years in Denmark (Pedersen et al. 2017). Higher mortality after hip fracture has been associated with a range of hospital factors (Kristensen et al. 2016, Sheehan et al. 2016) and patient factors in observational studies (Roche et al. 2005). Furthermore, variation in 30-day mortality after hip fracture surgery has been observed between Danish hospitals, but not fully explained (Kristensen et al. 2019).
背景与目的-了解髋部骨折手术后医院感染的关键驱动因素对于指导质量改进非常重要。因此，我们研究了髋部骨折手术后任何感染的风险，以及肺炎和败血症等亚组感染的风险。
方法-在这项全国性的基于人群的队列研究中，纳入了所有 2012 年至 2017 年期间因髋部骨折接受手术的年龄≥65 岁的丹麦患者。使用多层次泊松回归分析计算基于医院登记（医院治疗感染）和抗生素配药（社区治疗感染）的数据的术后感染风险。通过内类系数（ICC）和中位数风险比（MRR）评估医院间变异。
结果-医院治疗感染的风险为 15%。社区治疗感染的风险为 24%。调整后的风险在医院之间从 7.8%到 25%（医院治疗感染）和 16%到 34%（社区治疗感染）不等。ICC 表明，调整后变异的 19%归因于医院水平的医院治疗感染。社区治疗感染的 ICC 为 13%。MRR 显示，与风险最低的医院相比，平均患者在风险最高的医院接受医院治疗感染的风险增加了 2 倍。对于社区治疗感染，MRR 为 1.4。
解释-我们的研究结果表明，通过应用表现最好的医院的方法，可以减少 20%的感染。近五分之一的变异发生在医院层面。这表明需要采取更标准化的方法来避免髋部骨折手术后的术后感染。髋部骨折是老年人住院的主要原因。在丹麦，过去几年髋部骨折手术后的 30 天死亡率约为 10%（Pedersen 等人，2017 年）。在观察性研究中，髋部骨折后较高的死亡率与一系列医院因素（Kristensen 等人，2016 年，Sheehan 等人，2016 年）和患者因素相关（Roche 等人，2005 年）。此外，在丹麦医院之间观察到髋部骨折手术后 30 天死亡率的差异，但尚未完全解释（Kristensen 等人，2019 年）。
